Date | May Example question | Marks available | 13 | Reference code | EXM.2.AHL.TZ0.28 |
Level | Additional Higher Level | Paper | Paper 2 | Time zone | Time zone 0 |
Command term | Use, State, and Calculate | Question number | 28 | Adapted from | N/A |
Question
The random variable X is thought to follow a binomial distribution B (4, p). In order to investigate this belief, a random sample of 100 observations on X was taken with the following results.
An automatic machine is used to fill bottles of water. The amount delivered, Y ml, may be assumed to be normally distributed with mean μ ml and standard deviation 8 ml. Initially, the machine is adjusted so that the value of μ is 500. In order to check that the value of μ remains equal to 500, a random sample of 10 bottles is selected at regular intervals, and the mean amount of water, ˉy, in these bottles is calculated. The following hypotheses are set up.
H0: μ = 500; H1: μ ≠ 500
The critical region is defined to be (ˉy<495)∪(ˉy>505).
State suitable hypotheses for testing this belief.
Calculate the mean of these data and hence estimate the value of p.
Calculate an appropriate value of χ2 and state your conclusion, using a 1% significance level.
Find the significance level of this procedure.
Some time later, the actual value of μ is 503. Find the probability of a Type II error.
Markscheme
H0 : The data are B (4, p); H1 : The data are not B (4, p) A1
[1 mark]
Mean =1×32+…+4×14100 M1A1
= 1.8 A1
4ˆp=1.8⇒ˆp=0.45 M1A1
[5 marks]
The expected frequencies are
A1A1A1A1A1
The last two classes must be combined because the expected frequency for x = 4 is less than 5. R1
χ2=1729.15+32229.95+19236.75+32224.15−100 M2
= 18.0 A2
DF = 2 (A1)
Critical value = 9.21 A1
We conclude, at the 1% significance level, that X does not fit a binomial model. R1
Special case: award the following marks to candidates who do not combine classes.
χ2=1729.15+32229.95+19236.75+18220.05+1424.1−100 M2
= 39.6 A0
DF = 3 (A1)
Critical value = 11.345 A1
We conclude, at the 1% significance level, that X does not fit a binomial model. R1
[13 marks]
Under H0, the distribution of ˉy is N (500, 6.4). (A1)
Significance level = P ˉy < 495 or > 505 | H0 M2
= 2 × 0.02405 (A1)
= 0.0481 A1 N5
[5 marks]
The distribution of ˉy is now N (503, 6.4). (A1)
P(Type ΙΙ error) = P(495 < ˉy < 505) (M1)
= 0.785 A1 N3
[3 marks]